Active learning is a popular approach to teaching and learning that has gained traction through research on STEM educational improvement. There have been numerous university- and national/international-level efforts focused on transitioning courses from the lecture method to active learning. However, despite these large-scale changes, the active learning literature has not been assessed on its methodological rigor to ensure instructional recommendations are rooted in rigorous research studies. The purpose of the present review was to determine areas of strengths and areas in need of improvement and to provide specific recommendations on how to continue or improve active learning research to strengthen the respective literature base and increase confidence in results. We assessed the articles included in the Freeman et al. (PNAS, 111:8410–8415, 2014) meta-analysis as well as a random sample of more recent active learning articles (2015–2022) on 12 internal validity controls (i.e., control procedure used to prevent a threat to the internal validity of a study). Results indicated that there were high percentages of articles that did not meet each internal validity control. In fact, no articles from the Freeman et al. meta-analysis and no sampled 2015–2022 articles met each of the 12 internal validity controls. Therefore, the active learning literature contains numerous internal validity control issues that need to be addressed if we are to determine the extent to which active learning interventions are effective and if there are any boundary conditions for when particular active learning interventions are or are not effective.

主动学习是一种流行的教学方法，通过对 STEM 教育改进的研究而受到关注。许多大学和国家/国际层面都致力于将课程从讲座方法转变为主动学习方法。然而，尽管发生了这些大规模的变化，主动学习文献尚未对其方法论的严谨性进行评估，以确保教学建议植根于严格的研究。本次审查的目的是确定优势领域和需要改进的领域，并就如何继续或改进主动学习研究提供具体建议，以加强各自的文献基础并增强对结果的信心。我们评估了 Freeman 等人中包含的文章。（PNAS，111：8410-8415，2014）荟萃分析以及最近主动学习文章（2015-2022）的随机样本，涉及 12 个内部有效性控制（即用于防止内部有效性受到威胁的控制程序）的一项研究）。结果表明，有很高比例的文章不符合每项内部有效性控制。事实上，弗里曼等人没有发表任何文章。荟萃分析显示，2015-2022 年的样本文章均未满足 12 项内部有效性控制。因此，主动学习文献包含许多内部有效性控制问题，如果我们要确定主动学习干预措施的有效程度，以及特定主动学习干预措施何时有效或无效是否存在任何边界条件，则需要解决这些问题。